Spatial autocorrelation PreviousNext
Mplus Discussion > Structural Equation Modeling >
 Eric Le Tortorec posted on Thursday, January 05, 2012 - 10:18 am

I am conducting a SEM to analyse how the landscape structure at different scales (200-600m) affects the breeding success of the Eurasian treecreeper. For this analysis I am using an eight year dataset of breeding data collected from a set of about 200 nest boxes. A variogram analysis of the model residuals (obtained through SAS) showed that my response variable (number of fledged chicks) displays spatial autocorrelation. My question is: is there any way to take spatial autocorrelation into account in MPLUS?

Thank you very much for your help,
Eric Le Tortorec
 Bengt O. Muthen posted on Thursday, January 05, 2012 - 6:19 pm
There is currently not an explicit option for spatial analysis. You may want to have a look at the reference

Liu, Wall & Hodges (2005). Generalized spatial structural equation models. Biostatistics, 6, 539-557.

I don't know if Model Constraint can be used to capture any of those model features.
 Eric Le Tortorec posted on Friday, January 13, 2012 - 9:37 am
Dear Dr Muthén,

Thank you for your prompt answer.

As a follow up to my question, would it make sense to attempt to control for the spatial autocorrelation in the response variables before using them in the SEM? For example, I have considered using the residuals from an intercept-only glmm, which accounts for spatial autocorrelation in its variance-covariance structure, as the response variable.

Thank you for your help,
Eric Le Tortorec
 Bengt O. Muthen posted on Friday, January 13, 2012 - 11:18 pm
Maybe you want to contact the authors of the Biostat article and ask that question.
Back to top
Add Your Message Here
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Options: Enable HTML code in message
Automatically activate URLs in message